BFD can help with detection of link faults, which will then trigger faster failover. A point-to-point link does not mean a directly connected link. On a directly connected link, two devices connected to the same wire or fibre, link failure should be detected rapidly when the line protocol fails, thus forcing the interface down and traffic redirection. In networks where there are multiple transit L2 switches, the failure of a link in the transit may not bring down one of the OSPF peer's interface, the failure will effectively be hidden from the OSPF routers until the Hellos time out, and this can be very slow. In this case BFD will detect an indirect failure in the transit network much faster than OSPF and will trigger the switchover. That's the general idea, but in many Production networks, BFD is configured to underpin OSPF and BGP as a matter of course, as a necessary part of the operational process i.e. one doesn't configure it based upon knowledge of a switch's transit network, you just enable it as a best practice.
Thanks for the video. I had two routers in my scenario lo1-R1----lo-R2. I tried to redistribute connected on R1. On R2 I am not seeing lo1 as external LSA in the LSDB. I tried both with a route-map and without a route-map. Is it because loopback is not considered connected and is a local. Not sure how to get to loopback in. If I include the loopback in the area then it works. But that is not what I want. Thnx
There's no rules for this. The decision is based on operator preference. Factors that contribute are number of destination routes, number of next-hops to get to those destinations, number of devices that need to be maintained and whether there is the need for dynamic path learning. Static routing soon becomes unmanageable in two router network with, say, 20 different destinations that update regularly. Or even in a four router network, with multiple paths, and only a few destination subnets. If you play around with it in a lab with three routers, you'll soon see how much of a burden statics are. Better to just become comfortable with OSPF and deploy that.
Thanx for the video its quite helpful .. but I am not understanding how to advertise networks in this OSPF.. Like Below cisco config how here did. router ospf 1 passive-interface default no passive-interface GigabitEthernet1/1/1 no passive-interface GigabitEthernet1/1/2 network 10.162.2.1 0.0.0.0 area 0 network 10.162.12.1 0.0.0.0 area 0 network 10.162.13.1 0.0.0.0 area 0 network 10.162.14.1 0.0.0.0 area 0 network 10.162.15.1 0.0.0.0 area 0
You can. I created a spine-leaf topology with 3 spines and 4 leaf's. But not all features (like bfd) are functional. For more details look at release notes.
Appreciate the clear diction and understandable English. Goes a long way toward assimilation
Gracias from Texas!
Nice work
Excellent tutorial, it would be great if you could do a multi area and redistribution video. Thank you.
Should bfd also be enabled on point-2-point networks? is it faster with switching to another interface?
BFD can help with detection of link faults, which will then trigger faster failover. A point-to-point link does not mean a directly connected link. On a directly connected link, two devices connected to the same wire or fibre, link failure should be detected rapidly when the line protocol fails, thus forcing the interface down and traffic redirection.
In networks where there are multiple transit L2 switches, the failure of a link in the transit may not bring down one of the OSPF peer's interface, the failure will effectively be hidden from the OSPF routers until the Hellos time out, and this can be very slow.
In this case BFD will detect an indirect failure in the transit network much faster than OSPF and will trigger the switchover.
That's the general idea, but in many Production networks, BFD is configured to underpin OSPF and BGP as a matter of course, as a necessary part of the operational process i.e. one doesn't configure it based upon knowledge of a switch's transit network, you just enable it as a best practice.
Thanks for the video. I had two routers in my scenario lo1-R1----lo-R2. I tried to redistribute connected on R1. On R2 I am not seeing lo1 as external LSA in the LSDB. I tried both with a route-map and without a route-map. Is it because loopback is not considered connected and is a local. Not sure how to get to loopback in. If I include the loopback in the area then it works. But that is not what I want.
Thnx
When does one consider upgrading a medium network from static routing to ospfv2? BTW: another brilliant basic broadcast.
There's no rules for this. The decision is based on operator preference. Factors that contribute are number of destination routes, number of next-hops to get to those destinations, number of devices that need to be maintained and whether there is the need for dynamic path learning. Static routing soon becomes unmanageable in two router network with, say, 20 different destinations that update regularly. Or even in a four router network, with multiple paths, and only a few destination subnets. If you play around with it in a lab with three routers, you'll soon see how much of a burden statics are. Better to just become comfortable with OSPF and deploy that.
Found it! The command goes on the SVI..
Thanx for the video its quite helpful .. but I am not understanding how to advertise networks in this OSPF.. Like Below cisco config how here did.
router ospf 1
passive-interface default
no passive-interface GigabitEthernet1/1/1
no passive-interface GigabitEthernet1/1/2
network 10.162.2.1 0.0.0.0 area 0
network 10.162.12.1 0.0.0.0 area 0
network 10.162.13.1 0.0.0.0 area 0
network 10.162.14.1 0.0.0.0 area 0
network 10.162.15.1 0.0.0.0 area 0
You can redistribute from static or connected routes...
can i do that in GNS ?
You can. I created a spine-leaf topology with 3 spines and 4 leaf's. But not all features (like bfd) are functional. For more details look at release notes.
and vrrp video :c